Grouting in underground rock caverns by data mining

• Main Author: Chan, Wai Hong
• Other Authors: School of Civil and Environmental Engineering
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Civil engineering
• Online Access: http://hdl.handle.net/10356/77837
• Institution: Nanyang Technological University

Water seepage has always been a challenge to underground construction operations. Project delays and safety concerns are just some of the issues water seepage can bring about. Since groundwater tends to flow into newly excavated areas, this problem has become almost inevitable in any tunnelling projects. Grouting, which is a procedure of injecting a mixture of water and cement using pressure, is generally used to tackle water leakage. By plugging open spaces in the rock masses, grouting can reduce hydraulic conductivity and prevent excessive amounts of water seepage. However, establishing a suitable amount and pressure of grout, to ensure safe and economical use, is complicated. The aim of the report is to discover new grouting insights and potential relationships between grouting and other parameters using Artificial Neural Networks, a technique in data mining. Data sets from tunnel OT 0-1C of the Jurong Rock Caverns in Singapore are pre-processed and potential parameters are selected. These parameters undergo a preliminary analysis, which involves a correlation analysis that sieves out redundant parameters. Remaining parameters are fed into the neural networks to generate a predictive model. Subsequently, this model is analysed and cross-validated with data from another tunnel, OT 1-2. It was discovered that although the regression of the prediction model faired moderately well during the training and testing phase, the model produced poor results for cross-validation. This indicates that the model might not be accurate enough for general use and is only employable in the prediction of grouting parameters in tunnel OT 0-1C.